Plot spectroscopic data (a matrix) with ggplot2 in R

Question

I need to plot a spectroscopic data matrix, which rows are grouped by 2 factor variables, using ggplot2 (or lattice ) package in R as it has faceting capabilities.

Consider having a data frame DS with spectroscopic data (a matrix) DS$NIR from pls package:

library(pls)
data(gasoline)
DS <-gasoline

Let's add some grouping variables:

set.seed(0)
DS$Type <- as.factor(sample(c("Training set","Validation set","Others"),
                            nrow(DS),
                            replace = TRUE))

DS$Group <- cut(DS$octane,
                   breaks = c(80,86,88,90),
                   labels = c("Low","Medium","High"))

and look at data:

str(DS)

'data.frame':   60 obs. of  4 variables:
 $ octane: num  85.3 85.2 88.5 83.4 87.9 ...
 $ NIR   : AsIs [1:60, 1:401] -0.050193 -0.044227 -0.046867 -0.046705 -0.050859 ...
  ..- attr(*, "dimnames")=List of 2
  .. ..$ : chr  "1" "2" "3" "4" ...
  .. ..$ : chr  "900 nm" "902 nm" "904 nm" "906 nm" ...
 $ Type  : Factor w/ 3 levels "Others","Training set",..: 1 2 3 3 1 2 1 1 3 3 ...
 $ Group : Factor w/ 3 levels "Low","Medium",..: 1 1 3 1 2 1 3 3 3 3 ...

I need to plot every row of DS$NIR as a separate line. X axis values can be extracted by:

x <- as.numeric(gsub(" nm", "", dimnames(DS$NIR)[[2]]))

1. The line colors should depend on levels of factor Group .
2. The lines should be semi-transparent.
3. Every color group (ie every level of factor Group ) should have a solid opaque line, which indicates the average (or median) of the group.
4. Every level of factor Type should be plotted in a separate facet.

I found an example , how spectroscopic data is plotted, but yet it is too difficult for me to understand and adapt the code to my case.

Answer 1

You have two main problems with your data before plotting.

First, the NIR column is some weird matrix thing that doesn't play nicely with other functions. Let's fix that:

DS <- cbind(DS, as.data.frame(unclass(DS$NIR)))
DS$NIR <- NULL

Now, the data is wide, rather than long. Let's fix that with some dplyr and tidyr :

library(dplyr)
library(tidyr)    
graphdat <- DS %>% mutate(row = row_number()) %>%
                   gather(nm, value, -octane, -Type, -Group, -row) %>% 
                   mutate(nm = extract_numeric(nm))

Now it's easy to plot:

library(ggplot2)
ggplot(graphdat, aes(x = nm, y = value, group = row, color = Group)) + 
    geom_line() +
    facet_grid(Type~.)